Project Abstract Management of type 2 diabetes (T2D) in older patients is currently hampered by the lack of evidence needed to inform effective medical decision making. Older patients with diabetes are highly heterogeneous with regards to life expectancy, duration of diabetes, comorbidities, diabetic complications, disabilities, functional impairments, and treatment preferences. Current clinical guidelines all acknowledge this heterogeneity and support the concept of individualized diabetes care, but provide conflicting recommendations regarding how to individualize glycemic targets and medication regimens. These conflicts reflect the lack of evidence regarding the dynamic interactions between treatments, health status, and medical decision making in real-world clinical practice populations. While the patient perspective is a pillar of shared decision making, little is known about the variation in older patients' treatment preferences and experiences with current diabetes treatments, including self-management barriers and self-reported adverse drug events. Similarly, basic evidence is lacking on how the relationship between A1C and key outcomes varies by medical complexity and by classes of glucose-lowering medications. Moreover, busy clinicians lack practical, evidence-based tools to guide decision making regarding individualized targets and medications. Our overarching goal is to conduct observational research to provide evidence needed to inform safe and effective care of older adults with T2D. This research will be based on a well-characterized, multi-ethnic population of 145,894 patients ?65 years old with T2D from Kaiser Permanente Northern California. We will use the results of a NIA-funded survey (R56 AG051683) in an age-stratified, random sample of ~6,000 patients to characterize patient perspectives on treatment preferences, self-management barriers, and patient-related outcomes (e.g., hypoglycemia, falls, quality of life). In this proposed study, we will link survey responses to EMR-based exposures and outcomes (e.g., labs, medication prescribing and adherence, complications, mortality). This will allow us to characterize variation in older patient's experiences with and preferences for diabetes treatments and examine the relationships between patient preferences and self-management barriers with past or future glycemic control, medication use and outcomes (Aim 1). To inform efforts to establish appropriate glycemic targets, we will identify A1C levels that are associated with the lowest risk of key adverse diabetes outcomes (micro- and macrovascular complications, mortality, and hypoglycemia), stratified by health status, age, diabetes duration, and medication type (Aim 2). Finally, we will create a decision support tool to encourage the individualization of diabetes care, maximizing safety and minimizing overtreatment, in older adults with T2D (Aim 3). This effort includes developing and validating a contemporary mortality prediction model and integrating it with our existing hypoglycemia risk stratification model. The completion of the proposed studies will ensure that older patients and their providers have the clinical evidence and support necessary to make informed decisions for diabetes.